Riju - The Legend of Zelda (BotW and TotK) | Character Lora

If you enjoyed this model, why don't you check out my other works? I put a lot of effort into correctly recreating artist styles and making the occasional character LoRA.😊

You can also buy me a Coffee ☕ at https://ko-fi.com/stickyricky

V3.0

This should be the last version until we get a lot more high quality images of TotK Riju. It is mostly a refined version of v2. Most importantly lipstick now works and the totk version works well with other LoRA.

I recommend a weight of 0.8-1

For prompts check the model details and examples with further explanations below.

For one example I just threw all my style LoRAs into one XYZ plot. Some had their weight set a bit too high but overall I'm really happy with the flexibility. Let's hope that it works the same for different checkpoints.

Please leave a review if you enjoyed this Lora or have problems with it. I'd also be interested in how the model behaves with other base checkpoints so why don't you give it a try and throw some images down below!

V2.0

Now with a bit more experience in Lora creation I reworked it reusing a part of the original dataset and adding some images from Tears of the Kingdom. As before this model can be used for nsfw but doesn't enforce it to the same extent.

I recommend a weight of 0.8-1 with a weight of 1 being quite likely to include details like her crown without prompting and works with other artstyles. 0.8-0.9 should be used when you want to change her clothing or other details. As usual check out the XYZ plot for more details.

Some useful prompts are:

normal: long hair, hair rings

Even with the prompt hair rings the model will be more likely to generate two braids instead of the hair loop around her ear. The only thing that helps here is generating more images.

TotK: short hair, armor

details: blue lips, eyeshadow

Prompting for these details will push image generation towards more realistic and heavily shadowed versions. I'd recommend leaving this out, inpainting it or using heavily weighted lighting prompts.

For comparison I have redone some of the images from the previous version and included some examples of using different artstyles.

V1.0

This is my first functional attempt at a character LORA. Note that it was trained on Abyss Orange Mix 3 with images from gelbooru and is quite likely to yield NSFW results.

Props to holostrawberry for the guide to making LORAs:

[Guide] Make your own Loras, easy and free - colabs | Stable Diffusion Other | Civitai

For weights check the attached images. Generally 0.5-0.8 works and is pretty flexible with different artstyles. Hair color, jewelry and clothing details should be added depending on the used weight.